The Regulatory Effect of UL-16 Binding Protein-3 Expression on the Cytotoxicity of NK Cells in Cancer Patients

The activating immunoreceptor NKG2D (natural killer group 2, member D) and its ligands play important roles in the innate and adaptive immune responses. UL16-binding protein 3 (ULBP3), an NKG2D ligand, is overexpressed on certain epithelial tumor cells. In this study, we investigated the effect of ULBP3 expression on the cytotoxic activity of natural killer (NK) cells. ULBP3 were measured by flow cytometry analysis, immunohistochemistry, and time-resolved fluoroimmunoassay. The cytotoxicity of NK cells was determined with the lactate dehydrogenase release assay. We found that ULBP3 was overexpressed on tumor cell lines and tumor tissues. Serum from cancer patients, but not from healthy donors, contained elevated levels of soluble ULBP3 (sULBP3). Importantly, high expression of ULBP3 on the cell surface of tumor cells augmented NKG2D-mediated NK cell cytotoxicity. However, low levels of sULBP3 (<15 ng/ml) weakened the cytotoxicity of NK cells by decreasing NKG2D expression on NK cells. Further analysis showed that serum samples from most cancer patients (>70%) contained the low level of sULBP3. Our results demonstrate that tumor cells express surface and soluble ULBP3, which regulate NK cell activity. Thus, ULBP3 is a potential therapeutic target for improving the immune response against cancer.

The activating immunoreceptor NKG2D (natural killer group 2, member D) and its ligands play important roles in the innate and adaptive immune responses. UL16-binding protein 3 (ULBP3), an NKG2D ligand, is overexpressed on certain epithelial tumor cells. In this study, we investigated the effect of ULBP3 expression on the cytotoxic activity of natural killer (NK) cells. ULBP3 were measured by flow cytometry analysis, immunohistochemistry, and time-resolved fluoroimmunoassay. The cytotoxicity of NK cells was determined with the lactate dehydrogenase release assay. We found that ULBP3 was overexpressed on tumor cell lines and tumor tissues. Serum from cancer patients, but not from healthy donors, contained elevated levels of soluble ULBP3 (sULBP3). Importantly, high expression of ULBP3 on the cell surface of tumor cells augmented NKG2D-mediated NK cell cytotoxicity. However, low levels of sULBP3 (,15 ng/ml) weakened the cytotoxicity of NK cells by decreasing NKG2D expression on NK cells. Further analysis showed that serum samples from most cancer patients (.70%) contained the low level of sULBP3. Our results demonstrate that tumor cells express surface and soluble ULBP3, which regulate NK cell activity. Thus, ULBP3 is a potential therapeutic target for improving the immune response against cancer. N atural killer (NK) cells, components of the innate immune system, contribute to the elimination of virusinfected cells as well as to antitumor immune responses 1 . NK cell reactivity is guided by the principles of ''missing-self'' and ''induced-self,'' in which NK cells are activated by the downregulation or absence of major histocompatibility complex (MHC) I expression (''missing-self'') and/or by the stress-induced expression of ligands that bind activating NK receptors (''induced-self''). The balance of various activating and inhibitory signals determines whether NK cell responses are initiated [2][3][4][5] .
Among the activating NK receptors, NKG2D (natural killer group 2, member D) is particularly relevant for tumor cell recognition and killing. NKG2D is a C-type lectin-like activating receptor expressed on the cell surface of almost all NK cells, some cytotoxic CD8 1 ab T cells, NK T cells, and cd T cells, and a small subset of CD4 1 ab T cells [6][7][8] . NKG2D mediates NK cell activation by overcoming inhibitory signals from self recognition 9,10 . Malignant transformation induces the expression of NKG2D ligands (NKG2DL), as documented in a variety of human and mouse tumors. The activating immunoreceptor NKG2D endows cytotoxic lymphocytes with the capacity to recognize and eliminate malignant cells, and it plays a critical role in immune surveillance 11 . For example, NKG2DL-expressing tumor cells grafts were efficiently rejected, whereas parental NKG2D-ligand negative tumor cells formed tumors 12,13 . A distinctive feature of the NKG2D recognition system is that NKG2D can interact with a number of distinct ligands with affinities ranging from 4 to 400 nM [14][15][16] . The ligands recognized by NKG2D, which belong to distinct and relatively distantly related families, include major histocompatibility complex class-I related chain (MIC) A, MICB, and UL16-binding proteins (ULBPs) in humans 10,17 . NKG2DLs are generally not expressed on benign cells, but are induced by cellular stress, genotoxic stress, and infection 18,19 . The human ULBP proteins are widely expressed by various tumor types, including leukemia, and primary solid tumors [20][21][22] .
In addition to expressing NKG2DLs on their surface, tumors spontaneously release soluble ligands 23 . Soluble MICA secreted by tumor cells downregulated surface NKG2D expression on T cells to induce the functional impairment of anti-tumor immune effector cells, suggesting that shedding may reduce the expression of NKG2DLs on the tumor cell surface and contribute to tumor escape from immunosurveillance. Soluble MICA induced the internalization and lysosomal degradation of the NKG2D receptor in CD8 1 T and NK cells [24][25][26] , further reducing the efficiency of NKG2D recognition. Elevated serum levels of soluble MICA have been detected in patients with various types of cancer and may represent a diagnostic marker in patients with suspected malignancies 27,28 .
Unlike other NKG2DLs, ULBP3 has a moderate affinity for NKG2D. However, the regulatory function of ULBP3 in NK cells and its significance in cancer patients are largely unknown. In the present study, ULBP3 expression in several tumor cell lines and tumor tissue cells from common cancer types was analyzed. The effects of surface and soluble forms of ULBP3 on the interaction between tumor cells and NK cells were examined. Our results showed that ULBP3 regulated the activity of NK cells against tumors. Thus, ULBP3 provides a target for tumor immunotherapy.

Results
Elevated expression of ULBP3 in tumor cell lines and tumor tissues. To evaluate the distribution of the NKG2DL ULBP3 in tumor cells from common cancers, the surface expression of ULBP3 in SW620, K562, 7721, A549, and ECA109 cell lines was analyzed by flow cytometry (FCM) analysis. The colorectal cancer cell line CD133 2 SW620 expressed high levels (.50%) of ULBP3 (59.0 6 2.6%, n 5 3), and CD133 1 SW620 cells expressed moderate levels (20%-50%) of ULBP3 (22.0 6 1.4%, n 5 3). The liver cancer cell line 7721 also expressed a moderate level of ULBP3 protein (30.0 6 3.7%, n 5 3). However, surface ULBP3 protein was undetectable on the lung cancer cell line A549 and esophageal carcinoma cell line ECA109. The leukemic cell line K562, which does not express surface ULBP3, was used as a negative control ( Figure 1A). We then examined the expression of ULBP3 in different tumor tissues. In cancer patients with colorectal cancer (n 5 5), liver cancer (n 5 3), lung cancer (n 5 3), and gastric cancer (n 5 7), FCS indicated that ULBP3 expression was much higher in the tumor tissue than in the adjacent non-tumor tissue (ANTT). Representative dot graphs are shown in Figure 1B. Immunohistochemistry results also showed that ULBP3 expression was upregulated in colorectal (n 5 5), liver (n 5 3), lung (n 5 5), and gastric (n 5 5) cancer tissue ( Figure 1C). The results of RT-PCR analysis demonstrated that the ULBP3 gene was expressed in ULBP3-expressing tumor cell lines and in fresh tumor tissues from cancer patients, but not in ANTT ( Figure 1D).
Spontaneous release of soluble ULBP3 (sULBP3) from tumor cell lines and elevated serum sULBP3 in cancer patients. To determine whether tumor cells release sULBP3, we used a time-resolved fluoroimmunoassay (TRFIA) with a Eu 31 -labeled anti-ULBP3 monoclonal antibody. The detection limit of the TRFIA method for ULBP3-Fc was approximately 0.05 ng/ml ( Figure 2A). After 48 h of culture, sULBP3 levels in the supernatants of SW620, K562, 7721, A549, and ECA109 cells were measured. No sULBP3 was detectable in the culture supernatants of K562, A549, and ECA109 cells or in the culture supernatants of resting or activated NK cells. However, the levels of sULBP3 were detected in the supernatants of the tumor cell lines CD133 1 SW620 (3.6 6 0.08 ng/ml), CD133 2 SW620 (6.0 6 0.1 ng/ml), and 7721 (4.3 6 0.1 ng/ml) ( Figure 2B). To determine whether ULBP3 was released by human tumors in vivo, we analyzed serum sULBP3 levels in 116 patients with various malignancies and 48 healthy volunteers. Compared with healthy volunteers (1.5 6 0.1 ng/ml), the serum concentration of sULBP3 was significantly higher in cancer patients (colorectal cancer: n 5 45, 14.4 6 2.5 ng/ml, P , 0.001; gastric cancer: n 5 38, 8.4 6 1.1 ng/ml, P , 0.001; lung cancer: n 5 33, 9.6 6 1.4 ng/ml, P , 0.001) ( Figure 2C). Further analysis showed that 73.3% of colorectal cancer patients, 83.3% of gastric cancer patients, and 82.3% of lung cancer patients were distributed in the low level of serum ULBP3 (,15 ng/ml), as shown in Table 1.
Differences in the functional effects of surface ULBP3 and sULBP3 on NK cell activity. Next, we assessed the effect of ULBP3 cell surface expression on the lytic capacity of effector NK cells. CD133 1 SW620 (moderate expression of surface ULBP3), CD133 2 SW620 (high expression of surface ULBP3), and CD133 2 SW620 ULBP3-siRNA (no/ low expression of surface ULBP3; Figure 3A) tumor cells were used as target cells, and NK cells were freshly isolated from healthy volunteers for use as effector cells in cytotoxicity assays. With CD133 2 SW620 ULBP3-siRNA cells, the efficiency of NK-cell mediated killing was low when the effector cell to target cell (E5T) ratio was low (551 or 151) or high (1051). However, a high level of cytotoxicity was observed with CD133 2 SW620 cells, and a moderate level of cytotoxicity was found with CD133 1 SW620 cells. These results suggested that the amount of ULBP3 expressed on the cell surface of tumor cells affected the cytotoxicity of NK cells ( Figure 3B). In these experiments, blocking NKG2D on the surface of NK cells with a specific neutralizing monoclonal antibody strongly reduced the cytolytic activity of NK cells against the target cells at an E5T ratio of 1051, indicating that the NKG2D/NKG2DL pathway was involved in modulating the cytotoxic activity of NK cells ( Figure 3C).
To determine the effect of sULBP3 on the lytic capacity of NK cells, different concentrations of soluble recombinant ULBP3-Fc were added to the culture medium of tumor cells and NK cells, and cytotoxicity was measured. In the absence of surface ULBP3 expression on K562 and A549 cells, the addition of a low dose of sULBP3 (,15 ng/ml) reduced NK cell-mediated cytotoxicity against K562 cells at an E5T ratio of 1051. However, no significant increase in NK cell cytotoxicity was observed when high concentrations of sULBP3 (.15 ng/ml) were added to co-culture of NK and target cells ( Figure 4A and 4B). With 7721 and CD133 2 SW620 cells expressing moderate to high levels of surface ULBP3, similar results were observed ( Figure 4C and 4D). Then, we studied the effect of soluble ULBP3 on the lytic capacity of NK cells on different E5T ratios, we found that, firstly, in the non-expression of surface ULBP3 on K562 and A549 target cells, the presence of low dose of sULBP3 (1 ng/ml) caused a statistically reduction of NK cell cytotoxicity against target cells at E5T ratio of 551 and 1051. However, no significant increase of NK cell cytotoxicity was observed when high concentration of sULBP3 (30 ng/ml) was added in co-culture of NK and target cells ( Figure 4E and 4F). Secondly, in the moderate/high expression of surface ULBP3 on 7721 and CD133 2 SW620 target cells, the similar results were observed ( Figure 4G and 4H).
To confirm that sULBP3 regulated NK cell-mediated cytotoxicity, we incubated sULBP3-containing supernatants for 30 min in the absence or presence of an NKG2D-Fc fusion protein (5 ng/ml) before adding the supernatants to co-culture of NK cells and tumor cells. Neutralization of sULBP3 by NKG2D-Fc diminished the inhibitory effect of the sULBP3-containg supernatants on NK cell cytotoxicity against K562 ( Figure 5A), 7721 ( Figure 5B), and CD133 2 SW620 ( Figure 5C) cells, and no significant effects were detected in the NKp44-Fc control group.
To determine the mechanism by which sULBP3 affects the activity of NK cells, NK cells were pre-incubated with different concentrations of ULBP3-Fc for 6 h, and NKG2D expression was measured. When ULBP3-Fc was added to the culture medium, NKG2D expression on NK cells decreased, particularly with 1 ng/ml ULBP3-Fc. A representative graph depicting NKG2D expression is shown in Figure 5D, and combined data are shown in Figure 5E.  Figure 6A, and combined data are shown in Figure 6B. We then compared NK cytotoxicity in healthy individuals and in cancer patients with low (,15 ng/ml) or high (.15 ng/ml) levels of sULBP3. Freshly isolated NK cells from the peripheral blood mononuclear cells (PBMCs) of healthy individuals (n 5 3) and pre-therapy cancer patients (3 colorectal cancer, 2 lung cancer, and 3 gastric cancer) were used as effecter cells. K562 cells served as target cells. The cytotoxicity of NK cells from cancer Given that tumor cells generally express surface ULBP3 and sULBP3, we investigated whether an anti-ULBP3 antibody altered the cytotoxic activity of NK cells against ULBP3-expressing tumor cells. The lysis efficiency of the tumor cell lines CD133 2 SW620 ( Figure 6B) and 7721 ( Figure 6C) at an E5T ratio of 1051 was enhanced in the presence of anti-ULBP3 monoclonal antibody (B2-F1-F1). Furthermore, when an antibody against CD16 receptor was added, the effect was blocked, suggesting that the elevated activity of NK cells resulted from ADCC. Consistently, the downregulation of NK cytotoxicity by sULBP3 in K562 ( Figure 6D) or A549 ( Figure 6E) cells lacking ULBP3 was also reversed by B2-F1-F1. Together, these results suggested that B2-F1-F1 abrogated the sULBP3-induced inhibition of NK cell cytotoxicity and induced surface ULBP3-mediated ADCC to enhance the cytotoxicity of NK cells.

Discussion
By regulating the innate and adaptive immune response through ligation of NKG2D, NKG2DLs are thought to be important for tumor initiation and development. For example, NKG2DL expression has been associated with prognosis, immunoediting, and the control of an immunological checkpoint that relies on NKG2Dmediated immune responses during the epithelial-to-mesenchymal transition 29,30 . NKG2DL expression is of particular interest because previous experiments demonstrated that the NKG2D/NKG2DL system stimulates the immune surveillance of tumors 12,31 . Thus, NKG2DLs are potential targets for the development of novel immunotherapeutic anticancer regimens 11 . Although NKG2DLs such as MICA/B are widely expressed in many epithelial tumors and are associated with the regulation of NK and cd T cells, which play a key role in the elimination of tumor cells 32 , other NKG2DLs, such as ULBP1, ULBP2, and ULBP3, are not well characterized. In the present study, we demonstrated that ULBP3, like other NKG2DLs, was widely expressed on primary tumor cells and tumor cell lines. The NK cell cytotoxicity elicited by surface ULBP3-expressing tumor cells was stronger than that elicited by non-expressing tumor cells, suggesting that the expression of surface ULBP3 on tumor cells is directly related to the activity of NK cells against tumor cells.
Previous studies have demonstrated that the shedding of sULBPs can alter the density of surface-expressed NKG2DLs. The soluble molecules have systemic functions that affect NK cell reactivity against tumor cells 33 . Relative to their surface-expressed counterparts, sULBPs may mediate similar, partially different, or opposite effects. In the present study, we demonstrated that ULBP3 was spontaneously released in soluble form from ULBP3-expressing cells. In a high proportion of the serum samples obtained from cancer patients, sULBP3 levels were elevated, indicating that the release of sULBP3 was not an artifact of in vitro tumor cell culture but a feature of human tumors. However, our findings differ from those of previous reports describing the effect of soluble NKG2DL on NK cell activity. In our study, only a low dose of sULBP3 (,15 ng/ml) decreased NK cytotoxicity by downregulating NKG2D expression on NK cells; a high dose of sULBP3 did not inhibit NK cell activity. These findings that a low concentration of sULBP3 downregulated NK cell activity in vitro were consistent with the discovery that more than 70% of cancer patients had a low concentration (,15 ng/ml) of serum sULBP3. Furthermore, the percentage of infiltrating NK cells negatively correlated with ULBP3 expression. These results suggest that a low concentration of sULBP3 contributes to the hyporesponsiveness of the immune system in cancer patients by decreasing NK cell activity, and they explain a general aspect of immunological dysfunction in cancer patients.
Serum levels of soluble MICA/B are elevated in patients with various types of cancer. Soluble MICA/B may be a diagnostic marker in patients with malignancies 27,28 . Soluble ULBP2 has also been detected in the serum of some patients with malignancies 34 . We have developed a method for detecting serum ULBP3. Our data demonstrated that the concentration of sULBP3 was significantly higher in colorectal, lung, and gastric cancer patients than in healthy individuals, Furthermore, cancer patients released a significant amount of sULBP3 in vivo. In future studies, it will be important to determine whether the level of serum ULBP3 in cancer patients correlates with the progression and outcome of malignancies, in order to assess ULBP3 as a potential tumor marker.
When analyzing tumor tissue samples from cancer patients, we found a negative correlation between the percentage of infiltrating NK cells and the level of ULBP3 expression. Surface-expressed and       Immunohistochemistry. Colorectal cancer tissues, liver cancer tissues, lung cancer tissues, and gastric cancer tissues were obtained from the department of pathology in The Hospital Affiliated to Jiangsu University. Tissue sections were treated with 0.3% hydrogen peroxidase for 5 min and blocked for 30 min with normal goat serum at room temperature. Anti-ULBP3 monoclonal antibody (B2-FI-FI) (15200) was applied to the blocked sections and incubated overnight at 4uC. The sections were incubated for 30 min at 37uC with HRP-labeled goat anti-mouse IgG antibody (152000), and the signals were developed with diaminobenzidine tetrahydrochloride solution. The sections were viewed with an Olympus Ax-70 DMC Ie CCD camera connected to a PC monitor.
Cytotoxicity assay. NK cell cytotoxic activity against the tumor cells was determined by measuring the amount of lactate dehydrogenase (LDH) released from the target cells. A commercial LDH cytotoxicity kit (Beyotime) was used according to the manufacturer's instructions. The maximum LDH release was determined by lysing target cells for 30 min using the lysis buffer provided with the assay and measuring LDH in the culture medium. Absorbance for the colorimetric reaction was measured at a wavelength of 490 nm, with a reference wavelength of 655 nm, using a Model 550 microplate reader (Bio-Rad, Hercules, CA, USA). The specific lysis for each effector to target cell (E5T) ratio was calculated with the following formula: % specific lysis 5 [(experimental release 2 spontaneous release)/(maximum release 2 spontaneous release)] 3 100.
Time-resolved fluoroimmunoassay for measuring sULBP3 levels in the serum.
The wells of a 96-well plate were coated with the anti-ULBP3 monoclonal antibody B2-F1-F1 (1 mg/well) in coating buffer overnight at 4uC. Human recombination ULBP3-Fc was diluted to different concentrations for use as a standard. Serum samples and tumor cell culture suspensions were added at 100 ml/well and incubated at room temperature with shaking for 1 h and washed 6 times with buffer. Next, Eu 31 -labeled ULBP3 monoclonal antibody B4-C5-D11 (0.1 ng/well) was added, incubated at room temperature with shaking for 1 h, and washed 6 times with buffer. Accentuation buffer was then added at 100 ml/well, incubated with shaking for 0.5 h, and washed 6 times. The Eu 31 fluorescence intensity was measured using an AutoDELFIA 1235 immunoassay system (EG&G Wallac), and data were analyzed using the MultiCalc software.
Statistics analysis. The results are expressed as the mean 6 SEM. Comparisons between 2 groups were performed with Student's t-test. Differences among groups were assessed using one-way analysis of variance followed by the Tukey post hoc multiple comparisons test. P , 0.05 was considered statistically significant.